6.0 Existing Drainage Channels

Vancouver Port Authority Northwest Hydraulic Consultants Ltd/Triton Consultants Ltd. Roberts Bank Container Expansion File: 33863 Coastal Geomorphology Study – Appendix C - 46 - November 2004

6.0 EXISTING DRAINAGE CHANNELS

6.1 CHANNEL NETWORK

Figure C6-1 shows the present network of channels in the inter-causeway area. As described previously in Section 5.4, four main channel systems initially formed at the head of the original dredge basin in 1969 and appear to be sites of preferential channel formation when the ship turning basin was dredged in 1982. Channels 3 and 4 did not continue to develop, and are presently small and largely obscured by eelgrass. Channel 2, which was initially the largest channel, has been partially arrested by the crest protection structure and its overall development appears to have stalled. We have focused our attention on Channel 1 and Channel 2, which each have a network of smaller channels draining into a single trunk channel. The following sections describe these two channel systems in detail as well as the physical processes modifying them.

The channel system is characterised by a trunk (main) channel that extends from the crest protection structure upwards onto the tidal flats. Two main tributaries and a series of smaller channels deliver flow that drains from the tidal flats on the ebbing tide, to the upslope end of the trunk channel. The channels draining through and around the sediment lobe have a meandering planform and are connected in a dendritic pattern. Small channels seem to drain water off portions of the tidal flats on the ebbing tide, conveying flow into the main channels. At low tide, the larger channels have defined banks while the smallest channels appear as shallow depressions in the tidal flat surface.

The planform shape of the trunk channel is generally straight and the uppermost end terminates bluntly at a sandy bar deposit. The main trunk of Channel 1 has a width of approximately 90 m near the crest protection structure, and extends for a length of about 700 m. The trunk channel of Channel 1 splits at the crest protection structure, running parallel to the crest for a further 370 m before draining into the ship turning basin. The main trunk of Channel 2 has a width of approximately 40 m near the crest protection structure and extends 350 m up onto the tidal flats.

In planform, the channel systems are similar to ‘dendritic’ networks described by Howard (1967) and are analogous to a typical terrestrial-fluvial drainage network. This similarity allows the

Vancouver Port Authority Northwest Hydraulic Consultants Ltd/Triton Consultants Ltd. Roberts Bank Container Expansion File: 33863 Coastal Geomorphology Study – Appendix C - 48 - November 2004 adaptation of standard channel network descriptions (Strahler, 1957). Using this approach, the smallest tributary channels that could be identified on the orthophotos are designated as Order 1 streams. When two first order streams converge, the resulting larger channel is termed Order 2. Following this approach, the main trunk of Channel 1 is a fifth order stream and the main trunk of Channel 2 is a fourth order stream. Table C6-1 summarizes some other key network parameters, including drainage density (total stream length/total area), stream bifurcation ratio and average stream length. Figure C6-2 compares the drainage density of the two largest drainage networks with values reported on other basins throughout the world. The drainage density of Channel 1 is comparable to other basins and demonstrates that although the physical processes on Roberts Bank are very complex due to the varying effects of tides and waves, the overall channel network follows similar rules of behaviour and scaling relationships as conventional basins. The drainage density of Channel 2 plots near the lower limit of the graph, indicating the degree of channel development is much lower than other drainage networks. This is believed to be due to the influence of the crest protection structure which has partially arrested further channel development by preventing headcutting from occurring.

Table C6-1: Drainage Network Characteristics

Parameter Units Network 1 Network 2 Σ Length of Channels (m) 8,318 1,577 Area of Basin (m2) 2,887,036 143,960 Drainage Density (Σ -1 2.9 1.1 Length/Area) (km ) Average 1st Order (m) 25 45.3 stream length 2nd Order (m) 58 53.5 in each stream 3rd Order (m) 98 684 order 4th Order (m) 920 -- Average 1st Order (m) 16 18 stream width 2nd Order (m) 4 11 in each stream 3rd Order (m) 2 8 order 4th Order (m) 1 -- Order 1/Order 4.1 3.75 Bifurcation Order 2/Order 3.4 4 Ratio Order 3/Order 5-- 4

Vancouver Port Authority Northwest Hydraulic Consultants Ltd/Triton Consultants Ltd. Roberts Bank Container Expansion File: 33863 Coastal Geomorphology Study – Appendix C - 50 - November 2004

The network characteristics also provide a means for estimating the type of channels that can develop in defined drainage areas on the flats. For example, the drainage density and bifurcation ratios provide a means to determine the upper limit of Stream Order that can develop in a defined drainage area. For example, based on the observed network geometry in Channel 1, it can be shown that a drainage area of at least 0.25 km2 is required to generate a Third Order stream channel on the tidal flats (assuming a drainage density of 3 km/km2 and a bifurcation ratio of 4). Such channels would have a typical width of 4 m and a length of 100 m. Based on the present topography in the inter-tidal area, the approximate limiting elevation range for Third Order streams appears to be below El. +2m. Larger channels representative of Fourth Order or Fifth Order streams would not be expected to develop at higher elevation areas because the contributing drainage area is too small to sustain them. This finding is in agreement with the observed conditions on the higher portions of the tidal flats where only minor First Order and Second Order channels can be found.

6.2 CHANNEL CROSS-SECTION

The shape of the channel cross-section plays an important role in defining the drainage basin area. Figure C6-3 shows a typical cross-section surveyed during a low tide (RB/LB indicate right bank/left bank from the viewpoint of the ebbing flow). The section shows raised levees on either side of the channel, and a difference in water level elevation between the channel flow and the adjacent eelgrass beds. There is typically no vegetation in the channel or on top of the levees but outside the levees thick beds of eelgrass cover the tide flats. Eelgrass is sensitive to drying, so the beds thin shoreward, becoming sparse to non-existent above the 2 m elevation band. The absence of eelgrass within the channel zone is presumably because the in-channel flow velocities are too high and the tops of the levees are dry for a longer portion of the tidal cycle. At the outer margin of the levees, eelgrass is sparse and appears to be in the process of being buried by the levee as sediment is transported out of the channel and deposited on the eelgrass beds.

At lower tide levels the levees restrict flow from entering the channel from the surrounding eelgrass beds, except at specific points. Figure C6-4A shows a levee crevasse (breakthrough) with a small splay of sediment that has been deposited into the channel. Photo C6-4B shows a Vancouver Port Authority Northwest Hydraulic Consultants Ltd/Triton Consultants Ltd. Roberts Bank Container Expansion File: 33863 Coastal Geomorphology Study – Appendix C - 51 - November 2004 levee prograding into the eelgrass bed with a sparse growth of eelgrass at the outer margin that appears to be in the process of being overwhelmed by the depositing sediment.

Tidal-flat Channel Cross Section

1.0 RB LB Water Level 0.8 Water Level 0.6 0.4 0.2

Elevation (m) 0.0 0 20 40 60 80 100 120 Distance (m)

Figure C6-3: Cross Section of Tidal Channel

The levees described here are analogous to landforms found in other river systems, for example anastomosing rivers and birdsfoot type deltas. In both cases, overbank flow slows as it flows over the shallow banks and suspended sediment is deposited and trapped by vegetation growing along the channel margin. These levees slope gently from the channel bank into the flood basin outside the channel (Reinech and Singh, 1973, p.244). The term ‘flood basin’ describes the flood plain of an anastomosing river and the coastal bay of a delta. The analogous flood basin on the tidal flats at Roberts Bank is the eelgrass beds.

6.3 DISCHARGE MAGNITUDE

Estimates of the tidally-varying discharge in the drainage channels were made by (1) direct field measurements, (2) by two dimensional numerical modelling simulations and (3) by simplified tidal prism computations. Discharge and sediment transport are bi-directional, flowing up the tidal-flat slope as well as down. The discharge in the drainage channels is governed primarily by the tidal prism of the contributing drainage area, which in turn, is governed by the tidal range in the Strait of Georgia, modified by resistance and storage effects from eelgrass and local tidal flat topography. The range in flow magnitude and sediment transport is relatively limited due to the restricted range of the tide. Consequently, the upper limit to the flow magnitude in the drainage channels is approached several times each month. Measurements of discharges were made in Channel 1 periodically during the tidal cycle on three separate dates. These measurements

Vancouver Port Authority Northwest Hydraulic Consultants Ltd/Triton Consultants Ltd. Roberts Bank Container Expansion File: 33863 Coastal Geomorphology Study – Appendix C - 53 - November 2004 showed the maximum discharge in the channel reached 175 m3/s during an ebb tide and 135 m3/s during a flood tide. The measurements were made on May 8, 2004 during a Large Tide. Figure C6-5 shows a velocity profile across the channel on May 8th. Typical flow velocities in the Channel 1 trunk channel range between 0.5 and 1.0 m/s on both the ebbing and flooding tide.

6.4 OBSERVATIONS OF CHANNEL PROCESSES

The pattern of tidal flow over the flats was observed during several field visits. Field visits were typically made during the lowest tides of a given month because they would generally produce the greatest flow in the drainage channels. The following is a synopsis of these observations as they relate to channel formation and process.

On the falling tide, discharge across the flats follows the tide height, however there is a lag effect that varies from the upper to the lower portion of each drainage basin. At high tide, the tidal flats are inundated under several metres of water and the flow velocity over the flats is very low. Velocity increases gradually over the flats as the tide drops to reach a peak velocity at about the mid-point of the tidal range. The lag effect, induced by friction and temporary water storage on the flats, causes water to continue to drain from the tidal flats and into the drainage channels even as the tide reaches its minimum and begins to rise.

The velocities in the trunk channel are sufficient to develop dunes on the stream bed. Figure C6- 6 shows the bed profile the centreline of the main trunk channel during an ebbing tide. The profile was measured using an Odom Hydrographics echo sounder. Small dunes (typically 0.3 to 0.6 m high) are oriented in the seaward direction in the seaward end of the channel near the crest protection structure. However, near the head of the trunk channel the dunes were oriented landward, in spite of the ebbing tide.

Flow velocity and flow patterns at the crest protection structure undergo constant transition through the tidal range. At tide levels over two metres above the crest protection structure the water surface is uninterrupted by the presence of the rock. As the tide drops however, the water surface shows signs of the flow accelerating over the crest. At approximately one metre above the crest protection structure, water flows over the crest protection along its entire length with an average velocity of about 0.8 m/s. As the tide drops further, flow across the eelgrass beds

Vancouver Port Authority Northwest Hydraulic Consultants Ltd/Triton Consultants Ltd. Roberts Bank Container Expansion File: 33863 Coastal Geomorphology Study – Appendix C - 56 - November 2004 decreases and there are obvious jets of higher velocity flow issuing from the channel outlets and flowing across the crest protection structure, as well as a significant amount of flow being diverted along the structure and discharging at the southern end, through the Channel 1 outlet. These jets decrease as the tide continues to drop until at the lowest level of the tide the water surface in the turning basin is below the crest protection structure and the majority of flow from the tidal flat channels is diverted along the structure.

On the rising tide, water begins to flow upwards into the tidal flats in the lowest part of the channel paralleling the crest protection structure, while higher up on the tidal flats water continues to drain into the channels and flow down to meet the incoming tide. The most spectacular change in flow occurs as the rising tide overtops the large sandbar at the head of the Channel 1 trunk channel. Once the bar is overtopped the velocity increases rapidly and there is considerable transport of sand. Velocities of up to 0.7 m/s were measured in water depths of only 0.1 m. The direction of bar-surface ripples that were preserved during the ebbing tide were quickly reversed to reflect the up-flat direction of sediment transport, and antidunes forming on the bar surface indicated super-critical flow.

All evidence suggests a net shoreward transport of sediment in the vicinity of the large sandbar at the head of the Channel 1 trunk channel. We have confirmed through field observations that significant sediment transport occurs in this region on the rising tide.

6.5 COMPARISON WITH OTHER DRAINAGE CHANNELS

6.5.1 Purpose

In order to gain further insight into the mechanisms of channel formation on the tidal flats, we have examined other tidal drainage channels at Roberts Bank and Boundary Bay. These sites were chosen to be most representative of the channels draining into the ship turning basin because they are not overly influenced by freshwater runoff. These tidal channels are described here in terms of the physical characteristics of the channels and the specific physical characteristics of the tidal flats on which they are situated. The following is a list of channels that were examined for this study. The location of these channels is shown in Figure C6-7. Table C6-2 provides a comparative overview of the average dimensions of these tidal channels. Vancouver Port Authority Northwest Hydraulic Consultants Ltd/Triton Consultants Ltd. Roberts Bank Container Expansion File: 33863 Coastal Geomorphology Study – Appendix C - 57 - November 2004

• Channel 5: small channels draining from the northwestern portion of the inter-causeway tidal flats;

• Channel 6: Northwest BC Ferries channels, adjacent to and northwest of the BC Ferries causeway;

• Channel 7: BC Ferries boat basin channels, to the east of the BC Ferries causeway;

• Channel 8: BC Hydro channel, to the southeast of BC Ferries causeway; and

• Channel 9: Boundary Bay channel.

Table C6-2: Average Tidal Channel Dimensions on Roberts Bank and Boundary Bay

Parameter Tidal Channel Site 1256789 Width (m) 80 45 13 45 5 30 150

Ymax (m) 2.2 1.2 0.3 -- 0.5 -- --

Ymean (m) 1.5 1.1 0.2 -- 0.15 -- -- Top of Bank El. (m) 1.01.02.7------1.2 Thalweg El. (m) -1.1 -0.2 2.4 ------Length (m) 700 350 440 580 560 700 1,950 El. of 1st Bifurcation 1.0 1.0 -- 0.5-1.0 -- 1.0-1.5 1.8 ? (m) note: Ymax and Ymean are the maximum and mean depth of channel incision in the flats

6.5.2 Channel 5 – Northwest Corner of Inter-Causeway

This small channel drains the northwestern portion of the inter-causeway tidal flats. A pair of narrow, shallow channels has been visible on airphotos since as early as 1949 on the upper tidal flats in the northwest portion of the inter-causeway region. The channels appear to be formed by a combination of fresh and saltwater drainage from the upper marsh. There was very little change induced in these channels resulting from the construction of either the BC Ferries or Deltaport causeways. Until fairly recently the channels appeared to terminate naturally through dispersal of the flow at the 2.5 m contour (CD). Since the 1995 photos were taken however, it appears that a shallow channel has formed linking these two small channels to the larger system

Vancouver Port Authority Northwest Hydraulic Consultants Ltd/Triton Consultants Ltd. Roberts Bank Container Expansion File: 33863 Coastal Geomorphology Study – Appendix C - 59 - November 2004 of tidal channels upslope of the crest protection structure, providing continuous drainage from the top of the tidal flats to the seaward edge at the crest protection structure. It is not entirely clear if this apparent capture is due to up-flat growth of the dendritic channels, or if a shallow channel has existed for some time but remained undetected because of the clarity of the airphotos. In either case, the channel is typically only a few metres wide and less than 0.4 m deep. Interaction between the channel and vegetation on the tidal flats is not an important process with regards to channel formation.

6.5.3 Channel 6 – West Side of B. C. Ferry Terminal

This system of channels is located directly adjacent to the BC Ferries terminal and extends up the tidal flats past the eelgrass remediation site and along the causeway. It is composed of a main trunk and a pair of tributary channels. A rock weir, constructed across the mouth of the trunk channel in 1992, has resulted in a partial avulsion around the end of the barrier, similar to the partial avulsion around the end of the crest protection structure at the head of the turning basin.

The BC Ferries causeway and terminal have undergone three major periods of construction and expansion. The initial construction of the causeway and terminal in 1959 did not result in the formation of a channel that is visible on the airphotos. The first expansion of the terminal in 1976 included construction of three additional berths added to the southwestern end of the terminal. The 1975 airphotos show that prior to completion there was a fully developed system of channels to the north of the terminal that extended up along the causeway. These channels appear to have remained stable, with no discernible growth other than slight widening and lengthening, until 1991 when the terminal was again expanded. The 1990 airphotos were taken at a mid-tide and therefore the channels cannot be seen clearly. However, there does appear to be an up-flat progression of the channels by approximately 150 m. The second expansion of the terminal, completed in 1991, with the addition of a new parking area and eelgrass compensation site to the north of the terminal was partially constructed over the existing channels. This resulted in the channels being shifted to the north and the left bank tributary channel being completely buried. In 1992 a rock weir was constructed across the channel adjacent to the new parking area that forced the formation of a partial channel around the north end of the barrier. In Vancouver Port Authority Northwest Hydraulic Consultants Ltd/Triton Consultants Ltd. Roberts Bank Container Expansion File: 33863 Coastal Geomorphology Study – Appendix C - 60 - November 2004 the 12 years since, the channels do not appear to have increased appreciably in either width or lateral extent.

A slight rise in the inter-causeway tidal flats forms a divide between water draining into the dendritic channels at the crest protection structure and water draining southward into the channels paralleling the BC Ferries causeway and terminal. However, as there were no channels adjacent to the ferry terminal prior to the first expansion, it is likely that material was excavated from the tidal flats to construct the new berths. This is the most likely explanation for the very rapid formation of channels that have subsequently remained quite stable.

6.5.4 Channel 7 – East Side of B. C. Ferry Terminal

A deep trench was excavated in the tidal flats south of, and parallel to, the BC Ferries causeway to provide fill material for the causeway and terminal construction. The median depth in the upper part of the trench is between –2 and –2.5 m (CD), providing boaters a convenient spin-off benefit from the causeway construction.

A series of channels have formed around the upper margin of the boat basin in response to lowered base level. The channels between the trench and the causeway are quite small, reflecting the very short length of tidal flat in which they can form. There are three channels on the southeast side of the trench that are readily visible on the airphotos. They run roughly parallel to the shore, intersecting the trench at right angles. The largest of the three channels is the furthest to seaward and the mid-sized channel is the furthest shoreward. Field visits confirm that the channels appear to be formed by headward erosion of the tidal flats as flow drains from the eelgrass on the ebbing tide. Eelgrass grows in thick beds right to the channel margin and the bottoms of the channels are armoured with an oyster-shell pavement. Maximum width of the channels at the outlet is between 3 and 5 m, with depths of between 0.2 and 0.5 m.

The earliest post-construction photos (1960) do not show any evidence of channel formation into the trench. However, the airphoto record shows that by 1967 the largest of the channels had formed to nearly its present size, and by 1969 the other smaller channels had also formed. Since then, the channels do not appear to have expanded appreciably to the present (2004 airphotos). Vancouver Port Authority Northwest Hydraulic Consultants Ltd/Triton Consultants Ltd. Roberts Bank Container Expansion File: 33863 Coastal Geomorphology Study – Appendix C - 61 - November 2004

6.5.5 Channel 8 – BC Hydro Channel

This system of channels is situated approximately 1,800 m southeast of, and roughly parallel to the BC Ferries causeway. It is composed of a main trunk channel that is nearly straight and three tributary branches that drain off the tidal flats and into the head of the trunk channel. The seaward end of the trunk channel extends across the Canada-U.S. border. The main trunk is approximately 700 m long and the tributaries are up to 500 m long. The main trunk channel near the outlet is between 20 and 30 m wide and up to 1.5 m deep. The tributary branches are much smaller at only 5 m width and 0.75 m depth. Flow velocities in the main channel were not measured directly because the depth was too great for wading. Flow velocities at the mouth of the larger tributary channel were estimated approximately one hour after low tide but while water was still draining off of the upper tidal flats. They exceeded 1 m/s. The beds of the tributary channels were armoured with an oyster-shell pavement, while thick beds of eelgrass occupied the channel margins.

The BC Hydro Channel first appears in the airphoto record in 1959. However the tide level at the time the photo was taken is too high to see the channel clearly. By 1969 the channel appears on the airphotos in its present form though the channel has undergone some slight widening and lengthening in the interim.

The location of the trunk channel appears to correspond to the alignment of buried power cables that cross the tidal flats. It seems likely that disturbance from the installation of the cables has resulted in the formation of the channels. The tributary channels entering the upper end of the trunk channel have likely formed in response to the lowered base level in the trunk channel. These tributaries appear to have formed quite rapidly prior to 1969 and have subsequently remained fairly stable.

6.5.6 Channel 9 – Boundary Bay Channel

Although there are differences in the physical environment, it is worth examining the channels that have formed on the tidal flats of Boundary Bay, south of the City of Delta. The sediments in Boundary Bay were deposited during an earlier phase of the Fraser River delta and were subsequently isolated in the bay as the delta prograded westward (Kellerhals and Murray, 1969). Vancouver Port Authority Northwest Hydraulic Consultants Ltd/Triton Consultants Ltd. Roberts Bank Container Expansion File: 33863 Coastal Geomorphology Study – Appendix C - 62 - November 2004

Contemporary sediment sources in the bay include the unconsolidated sediments in the wave-cut cliffs at Point Roberts, erosion of sand from the marsh on the upper flats, and silts and clays delivered to the flats from the Nicomekl River. Human activity on the tidal flats dates back to the 1930s when oyster beds were cultivated as part of a flourishing industry (Kellerhals and Murray, 1969).

One of the large channels was selected in order to make a comparison to the tidal channels that have formed at Roberts Bank. Water drains from the upper and mid-flats through a large, straight channel that cuts across the flats perpendicular to the shoreline. The trunk channel is fed at the upper end by two large tributary channels that are in turn fed by a series of smaller channels forming a dendritic pattern. The upper channels are small, with a maximum width of less than 4 m. Accumulations of sediment in the form of sandbars appear at the junction of the trunk channel with the major tributaries.

Large-scale bedforms are visible on the tidal flats adjacent to the channel, suggesting that there is much higher sediment transport than at Roberts Bank, and offering a possible reason for the absence of eelgrass in this area. Furthermore, the overall elevation of the flats is higher than at Roberts Bank, particularly at the seaward end. Navigation charts produced by the Canadian Hydrographic Service show the elevation of the seaward margin to be 1.3 to 1.4 m CD within 300 m of the 0 m contour, while at Roberts Bank the elevation is shown to be 0.9 m for a similar offset from the 0 m contour. Thus the outlet portion of the channel is incised into the tidal flats by approximately 1.5 to 2 m.

Historical airphotos of Boundary Bay from the 1930’s and 1960’s show the general configuration of the channels has been remarkably stable. The long-term stability of the channels can also be inferred from their description in Kellerhals and Murray (1969) which shows their location on line drawing maps. Within the limits of mapping accuracy, the present location of this channel has not changed since 1969, though there have likely been some minor long-term adjustments in the channel. In the absence of evidence to the contrary, we have assumed, from geomorphic interpretation, that the channel is stable in the long-term. Vancouver Port Authority Northwest Hydraulic Consultants Ltd/Triton Consultants Ltd. Roberts Bank Container Expansion File: 33863 Coastal Geomorphology Study – Appendix C - 63 - November 2004

6.5.7 Summary

The following summary of observations is intended to provide background rationale to the conceptual model of channel response:

• The channel draining from the top of the tidal flats (Channel 5) predates all developments and has remained very stable. Of the channels discussed, it has formed at the highest elevation, therefore flow into and out of these channels occurs at the peak of the tidal curve when the rate of change of tide height, and therefore velocity, is at its lowest.

• The channels that drain into the top part of the BC Ferries Boat Basin and the northeast corner of the ship turning basin form as incised channels only and remain very small. There does not seem to be a sufficiently large tidal prism, reflected in the limited drainage basin area, to form larger channels.

• All of the channels that drain to the seaward edge of the tidal flats (Channels 1, 2, 6, 5, and 9) display a similar planform, with a straight main trunk channel fed by tributaries draining into the upslope end, with sandy bar deposits occurring at the confluence.

• Of the channel systems induced by trenches excavated parallel to the shore (Channels 6, 7, and 8), channel formation occurred very rapidly, often within a year, and then subsequently remained stable unless there was further human modification to the channels.

• The elevation at which the first bifurcation and sandy bar deposits occur in the tidal channels is consistently between 1.0 m and 1.5 m.

• The lowest and highest portions of the tidal flats appear to be dominated by ebb-tide flow processes, while the middle portion of the tidal flats appears to be dominated by processes occurring during flood tides. The morphology of the channels is reflected in the dominance of these two processes.